Analysis of Behavior Trajectory Based on Deep Learning in Ammonia Environment for Fish.

Area of Science:

• Aquaculture
• Animal Behavior
• Environmental Toxicology

Background:

• Ammonia is a byproduct of fish respiration and excretion in aquaculture.
• Elevated ammonia levels negatively impact fish health and survival.
• Current methods for monitoring fish behavior are often manual and labor-intensive.

Purpose of the Study:

• To investigate fish behavior under varying ammonia concentrations.
• To develop an automated system for early warning of abnormal fish behavior.
• To analyze the influence of ammonia on fish movement patterns.

Main Methods:

• Simulated different ammonia environments by adding ammonium chloride to water.
• Utilized deep learning for recognition and analysis of fish behavior trajectories.
• Employed three-dimensional reconstruction to generate fish spatial trajectories.

Main Results:

• Fish movement and vitality significantly decreased in elevated ammonia concentrations.
• Fish exhibited prolonged periods of stagnation in ammonia-containing water.
• Behavioral trajectory analysis revealed distinct patterns correlating with ammonia levels.

Conclusions:

• Deep learning-based behavior trajectory analysis offers a novel approach for monitoring fish health in aquaculture.
• The method provides a basis for early detection and warning of ammonia-induced stress in fish.
• This technique can be adapted for behavior analysis in other animal studies.